MEDINA:

Semi-Automatic Dublin Core to MPEG-7 Conversion

Research team Ralf Klamma (RWTH) Marc Spaniol (RWTH) Matthias Jarke (RWTH,  FIT)	Figure 1: MEDINA user interface
Students Monika Pienkos Dominik Renzel

Introduction

Knowledge creation processes in the cultural sciences have a discursive nature. The cultural scientists participating in these discourses build a community of practice. Reifications in information systems should support four concepts: (1) Content description by metadata, (2) coverage within standards, (3) repository technologies, and (4) platform independence of applications. By the combination of these concepts information systems support cultural scientists to extract and manage knowledge about high-level semantics of multimedia artifacts in open repositories with metadata annotations. Basic support for the latter can be achieved by a loose classification scheme as in Dublin Core, but with more sophisticated MPEG-7 description elements for time based media. MEDINA is a system for semi-automatic Dublin Core to MPEG-7 conversion to maintain media already annotated in Dublin Core.

Technical Aspects

The MPEG-7 Encoding of Dublin Core Information and Naming Application (MEDINA) has been specially designed for cultural science communities of researchers and students within the collaborative research center “Media and Cultural Communication”. In particular, it targets multidisciplinary research community in a project of the film studies. The community is physically distributed in the department of film studies at the Ruhr-Universität Bochum, the center in Cologne, University Bonn, and Munich. Its members have diverse backgrounds of education, e.g. film studies, history of art, graphical design and are on diverse levels of profession, i.e. full professors, research assistants, and students. Fig. 1 shows the user interface of MEDINA. It is divided into two parts. On the left hand side, the tabbed pane shows the user groups in MEDINA. Depending on the individual access rights the user can navigate through and add media to the corresponding media sets. On the right hand side, the tabbed pane offers four different functionalities: Thumb-preview of the media sets including a media-player, refinement of DC annotation by more sophisticated MPEG-7 descriptors, the MPEG-7 preview of the resulting file, and a temporal decomposition window containing the decompositions of a medium. MEDINA has been implemented as a platform independent Java application and offers two options to annotate multimedia artifacts. First, a new multimedia annotation can be created from scratch. Second, a DC annotation might be uploaded for an automated conversion to MPEG-7. Fig. 1 shows an uploaded DC document for a subsequent manual refinement possible with MPEG-7 descriptors.

Figure 2: Structural decomposition of MPEG-7 files in MEDINA

Technically, the conversion from DC to MPEG-7 is a two-stage process. First, the DC document is parsed internally in order to create a “normalized” DC document that contains the 15 elements (or a subset, in case not all of them have been used in the DC document) of the DC Metadata Element Set. The second step is the conversion itself. Here, the mapping between the DC Metadata Element Set [ISO01] and more sophisticated descriptors of the MPEG-7 multimedia metadata standard takes place. It is similar as proposed in [Hunt02], but has been refined for the particular needs of cultural science communities (cf. Appendix). For the sake of usability, additional MPEG-7 metadata concerning technical aspects of the medium such as visual encoding, frame rate, color space, etc. are extracted automatically. The DC elements are used for a loose metadata description offering a lot of semantic freedom in the scope of a more general metadata element. Thus, we’ve have subdivided the metadata file into three documents. A MEDINA “main” MPEG-7 metadata file with descriptors for technical information about the medium as well as most of the DC information. In addition, it contains references to two other MPEG-7 descriptions. The first type is a reference to “agents of action”. The second type is a reference to terms defined in MPEG-7 classification schemes that are used as a multi language “lexicon”. Thus, we are able to combine the more sophisticated role descriptors of MPEG-7 with DC elements. Fig. 2 shows the previously described structural decomposition of MPEG-7 documents in MEDINA. On the server side, we’ve set up an eXist-DB [Meie03] for the storage of the MPEG-7 metadata. Additionally, there is an affiliated FTP-server. The FTP-server is used for automated up- and download of multimedia artifacts by the community to the common repository.

References

[Hunt02]  J. Hunter. An Application Profile which combines Dublin Core
and MPEG-7 Metadata Terms for Simple Video Description.
http://www.metadata.net/harmony/video_appln_profile.html, Feb. 12 2002.

[ISO01]   ANSI/NISO Z39.85-2001: The Dublin Core Metadata Element Set.
http://www.niso.org/standards/resources/Z39-85.pdf, September 10 2001.

[Meie03]  W. Meier. eXist: An Open Source Native XML Database. In A. B. Chaudhri,
M. Jeckle, E. Rahm, and R. Unland, editors, Web, Web-Services, and Database
Systems, NODe 2002 Web and Database-Related Workshops, Erfurt, Germany,
October 7-10, 2002, Revised Papers, volume 2593 of LNCS, Springer-Verlag, Berlin
Heidelberg, pages 169 – 183, 2003.

Links

MPEG-7/21 Community

MECCA&MEDINA Community

Literature

M. Spaniol and R. Klamma: MEDINA: A Semi-Automatic Dublin Core to MPEG-7 Converter for Collaboration and Knowledge Management in Multimedia Repositories K. Tochtermann, H. Maurer (Eds.): Proceedings of I-KNOW '05, 5th International Conference on Knowledge Management, Graz, Austria, June 29 - July 1, 2005, J.UCS (Journal of Universal Computer Science) Proceedings, Springer, pp. 136-144.